In conventional cyclic transmission systems which are based upon the system of passing a token through a logical ring, the order of the stations for transmitting data is determined, for instance, according to the order of their addresses in a network management information table, and each of the stations transmits data after monitoring that the preceding station has finished data transmission.
However, according to such conventional transmission systems, when event data or incidental transmission data which arise at arbitrary timing coexist with cyclic data in each station, or when there are stations which are dedicated to the transmission of event data, each station is required to monitor the existence of any event data in the data transmitted from the preceding station, and is therefore required to be equipped with a monitoring timer. In particular, the determination of completion of a retransmission of event data after the occurrence of a transmission error is very time consuming.
According to a conventional method for controlling data transmission in a local area network (LAN), it has been known to provide a timer for determining the time period during which a token (transmission priority) can be held, and to allow data transmission from the time the token is received until the timer times out.
However, according to such a method for controlling the length of transmitted data, if the token is required to go entirely around the logical ring, it is necessary for the sum of the timer settings of all the stations to be equal to the time required for the token to be passed entirely around the logical ring, and the total sum of the timer settings must be checked before starting up the system. Further, to even more complicate the operation of the system, the timer settings of all the stations must be revised every time a station is either added to or removed from the logical ring, and the revision of the timer settings requires a considerable time.